Adverse prognostic factors for testicular cancer-specific survival: a population-based study of 27,948 patients.

The prognostic significance of age at testicular cancer (TC) diagnosis, socioeconomic status (SES), race, and marital status on TC-specific mortality is not well-characterized. In a cancer that is so curable, it is important to identify any influence that confers an increased risk of TC-specific mortality. Using multivariate cause-specific Cox regression models that accounted for competing risks, hazard ratios (HRs) were calculated for 10-year TC-specific mortality among 27,948 patients with TC reported to the Surveillance, Epidemiology and End Results program (1978 to 2006). Independent predictors were age at diagnosis, SES, race, marital status, extent of disease (EOD), calendar year of diagnosis, radiotherapy, and retroperitoneal lymph node dissection (RPLND). Compared with younger patients, diagnostic age 40+ was associated with increased mortality (seminoma, HR, 2.00, P < .001; nonseminoma, HR, 2.09; P < .001; most evident in metastatic disease, HR, 8.62; P < .001; HR, 6.35; P < .001, respectively). Unmarried men had two-to three-fold excess mortality compared to married men (HR, 2.97; P < .001; HR, 1.54; P < .001, respectively). Among nonseminoma patients, decreasing SES (P trend < .001) and nonwhite race (HR, 2.11; P < .001) increased mortality. Diagnosis after 1987 resulted in reduced mortality compared to earlier calendar years (HR, 0.58; P = .001; HR, 0.74; P = .001, respectively). Lack of RPLND was associated with seven-fold increase in death (P < .001). TC-specific mortality is doubled among US patients diagnosed with seminoma or nonseminoma after age 40, even when initial treatment and EOD are taken into account. Among men with nonseminoma, nonwhite race and lower SES also significantly increase TC-specific mortality. Additional research is needed, enabling the development of interventional strategies and preventive approaches, as applicable.